Switch TPv2 E2E tests to Cypress

[ocket8888]

This PR switches the TPv2 E2E testing suite from NightwatchJS to Cypress.

"Why?", you may ask. Because it took me all of two days to replicate all of our test suite in Cypress - that's how easy it is to use. For comparison: I have never even once been able to get the Nightwatch tests to run on my machine. With Cypress, the typings are provided by the library itself, testing on Chrome and Firefox took exactly zero time and effort to figure out, and the tests actually just do the things I want them to do without me having to spend countless hours needling them with the correct set of awaits and chained commands. Which is why most all of our tests do little more than check for the presence of some DOM elements. Maybe with Cypress, we could fix that.

tl;dr: Nightwatch is extremely hard to use and Cypress is extremely easy.

---

Which Traffic Control components are affected by this PR?

• Traffic Portal (experimental v2)

What is the best way to verify this PR?

Run the end-to-end tests. Which doesn't take very long and can be depended upon to actually work.

PR submission checklist

• This PR has tests
• This PR has documentation
• This PR doesn't need a CHANGELOG.md entry
• This PR DOES NOT FIX A SERIOUS SECURITY VULNERABILITY

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (adea91d) 28.84% compared to head (f232748) 74.35%.
Report is 2 commits behind head on master.

Additional details and impacted files

@@              Coverage Diff              @@
##             master    #7894       +/-   ##
=============================================
+ Coverage     28.84%   74.35%   +45.50%     
=============================================
  Files           602      117      -485     
  Lines         77445     5374    -72071     
  Branches         90      880      +790     
=============================================
- Hits          22342     3996    -18346     
+ Misses        53012     1310    -51702     
+ Partials       2091       68     -2023     

Flag	Coverage Δ
golib_unit	?
grove_unit	?
t3c_generate_unit	?
t3c_unit	?
traffic_monitor_unit	?
traffic_ops_integration	?
traffic_ops_unit	?
traffic_portal_v2	74.35% <ø> (?)
traffic_router_unit	?
traffic_stats_unit	?
unit_tests	74.35% <ø> (+48.68%)	⬆️
v3	?
v4	?
v5	?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[zrhoffman]

This is making the TPv2 RPM fail to build

From the PR of #7922:

npm ERR! Cypress cannot write to the cache directory due to file permissions

[ocket8888]

that's interesting. It shouldn't be attempting to compile or run the e2e tests to build the RPM.

[zrhoffman]

Reverted due to breaking tests in 4b6af89, please re-PR with tests fixed

[ocket8888]

Reverting something experimental because it affects the ability to build distributables for something that isn't distributed (because it's experimental) seems a bit extreme.

[zrhoffman]

If we don't want it to affect the default builds, then

• It should not be built when someone just runs ./pkg without specifying optional builds with -o, and
• It should not be included in the CDN-in-a-Box CI GHA workflow.

But because it is included in ./pkg by default and the CDN-in-a-Box CI GHA workflow, it broke tests for completely unrelated components, and that's why it was reverted.

[ocket8888]

• "It should not be built when someone just runs ./pkg without specifying optional builds with -o, and"

That should definitely never have been true. That's a bug in pkg, far as I'm concerned.

• "It should not be included in the CDN-in-a-Box CI GHA workflow."

Also should never have been true, but additionally: why is a test that depends on and explicitly uses something not run when that thing changes? Two bugs in the GHAs.

[zrhoffman]

AFAIAC, the bug that lead to the revert is that TPv2 is included in the CDN-in-a-Box CI GHA workflow but its paths don't trigger it. If the CDN-in-a-Box CI GHA had failed (due to the breaking changes in #7894), it would not have gotten merged and would not need to be reverted.